Asteria And Moonvalley Partner To Bring First Ethical AI Model To Hollywood. In a string of important announcements this week, Bryn Moser, founder of RYOT, a studio for immersive content which was sold to Verizon, founder and CEO of venture-funded, ocscar-nominated, non-fiction streamer, XTR, has finally unveiled the shape of his AI Production company, housed in the storied old Sennett Studios in LA’s Silverlake neighborhood. XTR, whose AI spin off is called Asteria (the Greek goddess who was “of the stars”), purchased a pioneering animation research company, Late Night Labs, earlier this year. Asteria gave some no-strings grants to AI filmmakers to foster community, creativity and collaboration. I didn’t understand the commercial purpose of this until Asteria annoucned the Moonvalley partnership. At the same time, Asteria Announced It Landed Filmmaker Paul Trillo As Creative Partner. Trillo and Asteria will likely be the first to produce content with the new Moonvalley model. If it’s really good, the world is going to beat a path to Moser and Moonvalley’s door. And a tidal wave of fast followers will come pouring in with their own ‘clean’ models.
Google unveiled Veo 2, its latest AI-powered video generator. They claim it outperforms OpenAI’s SORA model in audience satisfaction tests. Veo 2 produces complex, multi-scene videos guided by user instructions, showcasing stronger thematic consistency, improved quality, and better coherence. It employs advanced diffusion techniques, and incorporates human feedback to refine its outputs, aiming to streamline content creation workflows for creators and marketers. By reducing complexity and production times, Veo 2 seeks to democratize high-quality video generation. Although still in early research stages, Veo 2 signals Google’s dedication to advancing generative AI tools and reshaping how video content is produced.
Higgsfiled Launches ReelMagic, Which Generates 60-second Videos. ReelMagic is a comprehensive story-creation tool that turns a single prompt into a video narrative. Users can start from a simple logline to generate scripts, keyframes, and animated characters automatically. With integrated fandom support, including well-known universes, and customizable Lora model uploads, creators can easily incorporate recognizable characters or design unique ones. The platform unites top-tier animation, text, and voiceover models—Higgsfield Flux, Kling, Minimax, Runway, and ElevenLabs—within a single interface, eliminating the need to switch between services. Batch generation accelerates production, allowing multiple keyframe variations and simultaneous video outputs. Voiceover assignment is streamlined directly into the timeline. Finally, snapshots enable quick collaboration, versioning, and inspiration by duplicating or building on existing projects. The company was founded by Alex Masharbov, the former head of AI for Snap. It’s still in closed beta but it’s epic promise is clear.
Pika Releases 2.0 Gen-AI Model. This new model gives you the ability to prompt for specific elements in a scene, such as a certain outfit, person, or environment.
Krea is putting every video model in one place.
Minimax recently launched a new AI video model that specializes in animated visuals. It often tries to change it to live action or does some type of weird hybrid thing. This new model provides much more consistency.
Project Odyssey, from Civit.ai is hosting its second annual AI Filmmaking competition, with $70,000 in prizes from some of the most notable AI companies in the world, including Kling, ComfyUI, Viggle, and so many more.
More Cinematic AI. Every one of these extraordinary shorts was made by an entertinment or creative industry professional.
